1+ #
2+ # CLOUDERA APPLIED MACHINE LEARNING PROTOTYPE (AMP)
3+ # (C) Cloudera, Inc. 2024
4+ # All rights reserved.
5+ #
6+ # Applicable Open Source License: Apache 2.0
7+ #
8+ # NOTE: Cloudera open source products are modular software products
9+ # made up of hundreds of individual components, each of which was
10+ # individually copyrighted. Each Cloudera open source product is a
11+ # collective work under U.S. Copyright Law. Your license to use the
12+ # collective work is as provided in your written agreement with
13+ # Cloudera. Used apart from the collective work, this file is
14+ # licensed for your use pursuant to the open source license
15+ # identified above.
16+ #
17+ # This code is provided to you pursuant a written agreement with
18+ # (i) Cloudera, Inc. or (ii) a third-party authorized to distribute
19+ # this code. If you do not have a written agreement with Cloudera nor
20+ # with an authorized and properly licensed third party, you do not
21+ # have any rights to access nor to use this code.
22+ #
23+ # Absent a written agreement with Cloudera, Inc. ("Cloudera") to the
24+ # contrary, A) CLOUDERA PROVIDES THIS CODE TO YOU WITHOUT WARRANTIES OF ANY
25+ # KIND; (B) CLOUDERA DISCLAIMS ANY AND ALL EXPRESS AND IMPLIED
26+ # WARRANTIES WITH RESPECT TO THIS CODE, INCLUDING BUT NOT LIMITED TO
27+ # IMPLIED WARRANTIES OF TITLE, NON-INFRINGEMENT, MERCHANTABILITY AND
28+ # FITNESS FOR A PARTICULAR PURPOSE; (C) CLOUDERA IS NOT LIABLE TO YOU,
29+ # AND WILL NOT DEFEND, INDEMNIFY, NOR HOLD YOU HARMLESS FOR ANY CLAIMS
30+ # ARISING FROM OR RELATED TO THE CODE; AND (D)WITH RESPECT TO YOUR EXERCISE
31+ # OF ANY RIGHTS GRANTED TO YOU FOR THE CODE, CLOUDERA IS NOT LIABLE FOR ANY
32+ # DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, PUNITIVE OR
33+ # CONSEQUENTIAL DAMAGES INCLUDING, BUT NOT LIMITED TO, DAMAGES
34+ # RELATED TO LOST REVENUE, LOST PROFITS, LOSS OF INCOME, LOSS OF
35+ # BUSINESS ADVANTAGE OR UNAVAILABILITY, OR LOSS OR CORRUPTION OF
36+ # DATA.
37+ #
38+
39+ #
40+ # CLOUDERA APPLIED MACHINE LEARNING PROTOTYPE (AMP)
41+ # (C) Cloudera, Inc. 2024
42+ # All rights reserved.
43+ #
44+ # Applicable Open Source License: Apache 2.0
45+ #
46+ #
47+ # This code is provided to you pursuant a written agreement with
48+ # (i) Cloudera, Inc. or (ii) a third-party authorized to distribute
49+ # this code. If you do not have a written agreement with Cloudera nor
50+ # with an authorized and properly licensed third party, you do not
51+ # have any rights to access nor to use this code.
52+ #
53+ # Absent a written agreement with Cloudera, Inc. ("Cloudera") to the
54+ # contrary, A) CLOUDERA PROVIDES THIS CODE TO YOU WITHOUT WARRANTIES OF ANY
55+ # KIND; (B) CLOUDERA DISCLAIMS ANY AND ALL EXPRESS AND IMPLIED
56+ # WARRANTIES WITH RESPECT TO THIS CODE, INCLUDING BUT NOT LIMITED TO
57+ # IMPLIED WARRANTIES OF TITLE, NON-INFRINGEMENT, MERCHANTABILITY AND
58+ # FITNESS FOR A PARTICULAR PURPOSE; (C) CLOUDERA IS NOT LIABLE TO YOU,
59+ # AND WILL NOT DEFEND, INDEMNIFY, NOR HOLD YOU HARMLESS FOR ANY CLAIMS
60+ # ARISING FROM OR RELATED TO THE CODE; AND (D)WITH RESPECT TO YOUR EXERCISE
61+ # OF ANY RIGHTS GRANTED TO YOU FOR THE CODE, CLOUDERA IS NOT LIABLE FOR ANY
62+ # DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, PUNITIVE OR
63+ # CONSEQUENTIAL DAMAGES INCLUDING, BUT NOT LIMITED TO, DAMAGES
64+ # RELATED TO LOST REVENUE, LOST PROFITS, LOSS OF INCOME, LOSS OF
65+ # BUSINESS ADVANTAGE OR UNAVAILABILITY, OR LOSS OR CORRUPTION OF
66+ # DATA.
67+ #
68+
69+ import os
70+ import socket
71+
72+ def validate ():
73+ print ("Validating environment variables..." )
74+ access_key_id = os .environ .get ("AWS_ACCESS_KEY_ID" , None )
75+ secret_key_id = os .environ .get ("AWS_SECRET_ACCESS_KEY" , None )
76+ default_region = os .environ .get ("AWS_DEFAULT_REGION" , None )
77+ document_bucket = os .environ .get ("S3_RAG_DOCUMENT_BUCKET" , None )
78+
79+ caii_domain = os .environ .get ("CAII_DOMAIN" , None )
80+
81+ # 1. if you don't have a caii_domain, you _must_ have an access key, secret key, and default region
82+ if caii_domain is None :
83+ if access_key_id is None or secret_key_id is None or default_region is None :
84+ print ("ERROR: Using Bedrock for LLMs/embeddings; missing required environment variables: AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_DEFAULT_REGION" )
85+ exit (1 )
86+ else :
87+ try :
88+ socket .gethostbyname (caii_domain )
89+ print (f"CAII domain { caii_domain } )
90+ except socket .error :
91+ print (f"ERROR: CAII domain { caii_domain } )
92+ exit (1 )
93+
94+ # 2. if you have a document_bucket, you _must_ have an access key, secret key, and default region
95+ if document_bucket is not None :
96+ if access_key_id is None or secret_key_id is None or default_region is None :
97+ print ("ERROR: Using S3 for document storage; missing required environment variables: AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_DEFAULT_REGION" )
98+ exit (1 )
99+
100+ if caii_domain is not None :
101+ print ("Using CAII for LLMs/embeddings; CAII_DOMAIN is set" )
102+
103+ else :
104+ print ("Using Bedrock for LLMs/embeddings; AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, and AWS_DEFAULT_REGION are set" )
105+
106+ if document_bucket is not None :
107+ print ("Using S3 for document storage (S3_RAG_DOCUMENT_BUCKET is set)" )
108+ else :
109+ print ("Using the project filesystem for document storage (S3_RAG_DOCUMENT_BUCKET is not set)" )
110+ # TODO: verify that the bucket prefix is always optional
111+
112+ validate ()
0 commit comments